ce-humanizelisted

# Humanize **Invocation input:** Text, a file path, or tone/audience context — plus what to rewrite: $ARGUMENTS Rewrite the provided prose so it reads like it was written by a competent human with real context and judgment — natural, specific, and credible — **without changing what it says**. This is an editorial transform, not a content generator: it never invents facts, never adds fabricated experience, and never claims the output was human-written or tries to defeat AI-detection tooling. It is **repo-agnostic** and **not** part of the plan/spec/implement pipeline — it consumes and produces none of those artifacts. Use it on any prose, in any repo, whether or not the repo uses spec-driven development. ## Runtime Inputs - **What to rewrite (required):** raw text pasted in, a **file path**, or a selection already in the conversation. If none is supplied, ask for it before doing anything else. - **Context (optional):** intended audience, medium, tone (`professional`, `conversational`, `executive`, `technical`, `academic`, `social`), a length bound, or "give me options". Inferred when unstated — never invented into the content itself. ## Execution Contract 1. **Meaning is invariant.** Preserve the source's claims, intent, technical accuracy, and required terminology exactly. Rewriting changes *how* it reads, never *what it asserts*. 2. **No fabrication.** Never add facts, data, quotes, citations, credentials, emotions, events, or personal/lived experie